Welcome to the 3Ranker portal

This is the current project page for the 3Ranker project, a collaboration between the Karolinska Institute, SYRCLE and TenWise and was financed by the Swedish Fund for Research Without Animal Experiments.

3R stands for the 3 ways of implementing alternatives to animal experiments by means of :

  • Replacement
  • Avoiding or replacing the use of animals in areas where they otherwise would have been used.
  • Reduction
  • Minimising the number of animals used consistent with scientific aims.
  • Refinement
  • Minimising the pain, suffering, distress or lasting harm that research animals might experience.

In this project we develop a search algorithm (3Ranker) for 3R papers that is based on Artificial Intelligence. 3Ranker currently works for 35 million abstracts that are indexed in MEDLINE.

The portal has 3 main tabs for you to discover :

  • Search
  • The search tab enables you to search across all abstracts indexed in MEDLINE from any given input term. The search results are then scored using the 3Ranker algorithm and the 200 abstracts with the highest score are displayed in the results table for you to further analyse.
  • Curated Abstracts
  • The curated abstracts tab contains a table displaying all abstracts that were rated as 3R-related by certified experts.
  • Reference Scoring
  • The reference scoring tab is used to actively review new abstracts in terms of whether they are related to 3R. This is a continuous process to keep on improving the 3R algorithm.

Free Search

Use the search term for a keyword search in PubMed. The analysis will be done on the 5000 most recent abstracts.





The data table to the right shows all curated abstracts which are related to the 3R principle.
These abstracts have undergone thorough screening by certified experts and are displayed here to enable navigation throughout the 3R related literature. The search bar and interactive column buttons let you filter on abstracts that you find relevant.

Curated Abstracts



This tool uses a machine learning algorithm that is trained via reference scoring to discriminate between scientific abstracts that are about animal replacement (and even broader: the 3Rs) and those that are not. This saves researchers a lot of time in finding out if animal replacement is possible because such a check within research has now become a requirement by EU regulations.

How does the assessment work?

In order to start the assessment, reviewers have to select their assigned Reviewer number in the left panel. This number has been sent in a separate email. They always have to select this reviewer number when scoring abstracts for the 3Ranker. The number of ranked abstracts are shown next to the reviewer numbers.

In the middle of the page abstracts with title are shown. Each abstract is subject for assessment solely based on the provided title and abstract text. A new abstract is shown after the red, orange and green button is clicked. Reviewers click red if they are sure this has nothing to do with 3R (animal replacement), orange if they are not sure, and click green if they are sure the article deals with 3R.

Definition of 3R

3R is in this case defined in line with the EU Directive 2010/63EU’s definition of Replacement: “A scientifically satisfactory method or testing strategy, not entailing the use of live animals, shall be used instead of a procedure. This implies any scientific method employing non-sentient material replacing or avoiding using conscious whole living animals in experiments where they would otherwise have been used."


  • An in vitro study (eg. using human or animal cells) that could replace/avoid a study in live animals.
  • An in silico study (eg. using a computer of computer similation only) that could replace/avoid a study in live animals.
  • An in vitro or in silico study using adverse outcome pathways that could replace/avoid a study in live animals.
  • Studies that involve the use of certain animal species that are currently in the scientific thinking not considered capable of experiencing suffering, like nematode worms, social amoebae, Drosophila and immature forms of vertebrates.
  • In general human studies will not be considered Replacement, as they would have been performed anyway.

Please note the following

  • An abstract may appear more than once, as we use a pool of ~1,000 abstracts, but please assess it every time.
  • Please base your assessment on the provided text of the abstract and title only!
  • The input is kept anonymous.

Additional questions?

Do you encounter any technical problems? Please let us know: nils.hijlkema@tenwise.nl/brun.ulfhake@ki.se

Reviewer Access

Please login to get full access as a reviewer on the contribution page. If you are interested in being a certified reviewer, please contact us.